Before integrating
- Confirm a public model name in the catalog.
- Check status, context and feature support.
- Run a small non-streaming request.
- Verify usage and pricing.
- Only then enable production traffic.
Why the page remains useful when no model is listed
It gives developers a truthful availability check and directs alternative searches to a separate comparison page instead of fabricating a model route.
Live model catalog
| Model | Input / 1M tokens | Output / 1M tokens | Context | Status |
|---|---|---|---|---|
| DeepSeek-V4.1-Flash-VIP | CNY 0.8 | CNY 3.2 | 1.0M | Available |
| deepseek-v4-1-flash-260910 | CNY 0.8 | CNY 3.2 | 1.0M | Available |
| deepseek-v4-1-flash1017 | CNY 0.8 | CNY 3.2 | 1.0M | Available |
| deepseek-v4.1-flash-VIP | CNY 0.8 | CNY 3.2 | 1.0M | Available |
| DeepSeek-V4-Pro | CNY 3.6 | CNY 10.8 | 1.0M | Available |
Common questions
Is DeepSeek available right now?
Check the live table on this page. No matching row means this site is not claiming a currently callable public DeepSeek model.
Can I use the OpenAI SDK if a DeepSeek model appears?
Use the public model name and compatible endpoint, then verify model-specific behavior in a test request.